Thursday, November 30, 2023
HomeBiologyGender imbalances amongst top-cited scientists throughout scientific disciplines over time by the...

Gender imbalances amongst top-cited scientists throughout scientific disciplines over time by the evaluation of practically 5.8 million authors


We evaluated how the gender composition of top-cited authors inside totally different subfields of analysis has advanced over time. We thought-about 9,071,122 authors with not less than 5 full papers in Scopus as of September 1, 2022. Utilizing a beforehand validated composite quotation indicator, we recognized the two% top-cited authors for every of 174 science subfields (Science-Metrix classification) in 4 separate publication age cohorts (first publication pre-1992, 1992 to 2001, 2002 to 2011, and post-2011). Utilizing NamSor, we assigned 3,784,507 authors as males and a pair of,011,616 as ladies (for 36.1% gender task unsure). Males outnumbered ladies 1.88-fold amongst all authors, lowering from 3.93-fold to 1.36-fold over time. Males outnumbered ladies 3.21-fold amongst top-cited authors, lowering from 6.41-fold to 2.28-fold over time. Within the youngest (post-2011) cohort, 32/174 (18%) subfields had > = 50% ladies, 97/174 (56%) subfields had > = 30% ladies, and three subfields had = <10% ladies among the many top-cited authors. Gender imbalances in creator numbers decreased sharply over time in each high-income international locations (together with the USA of America) and different international locations, however the latter had little enchancment in gender imbalances for top-cited authors. In random samples of 100 ladies and 100 males from the youngest (post-2011) cohort, in-depth evaluation confirmed that the majority had been presently (April 2023) working in tutorial environments. 32 ladies and 44 males had some college appointment, however solely 2 ladies and a pair of males had been full professors. Our evaluation reveals giant heterogeneity throughout scientific disciplines within the amelioration of gender imbalances with extra outstanding imbalances persisting amongst top-cited authors and sluggish promotion pathways even for the most-cited younger scientists.


Gender disparities have been very outstanding in science throughout a number of dimensions together with recruitment, tenure, funding, authorship, and quotation affect [15]. A few of these disparities could also be diminishing over time, however the tempo of change varies throughout scientific fields, settings, and international locations. For instance, an evaluation report [6] has documented the lowering hole between the numbers of feminine and male authors throughout science through the years, however whereas the hole has virtually disappeared in Argentina, it continues to be very giant in Japan. Furthermore, in the identical evaluation [6] when medical subfields are examined, the inequality continues to be very sturdy towards ladies within the fields of surgical procedure and radiology and imaging, whereas ladies authors outnumber males presently in fields comparable to infectious illnesses, fertility, and public well being.

Quotation affect specifically is a key coinage within the scientific tutorial enterprise and there’s proof that citations are misused and gamed [7]. The significance of citations each as a way of educational energy in addition to a marker and promoter of inequalities could also be most outstanding among the many most-cited scientists and may additionally have an effect on tutorial profession trajectories. Gender imbalance in scientific careers could also be pushed be a number of advanced forces [8,9], however publications and citations could also be key widespread mediators. Variations in quotation counts might replicate distinction within the variety of publications and/or within the citations obtained per publication for authors of various genders [6]. Earlier evaluation has proven [6] that, total, ladies are inclined to publish fewer papers than males and that the field-adjusted variety of common citations obtained is modestly bigger when first creator is a person relatively than a lady.

Right here, we purpose to make use of complete publication and quotation information that cowl all science by the Scopus database [10] in an effort to consider how gender disparities have modified over time within the cohorts of the most-cited scientists, throughout every of the 174 subfields of science [11]. Prime-cited scientists are a choose group that’s the most influential throughout science, and any gender biases on this group are more likely to have main repercussions for science at giant. The obtainable information that we’ve got compiled permit us to analyze cohorts of scientists in accordance with their publication age (i.e., what number of years they’ve been energetic publishing scientific work). For scientist cohorts of various publication age, we determine the two% top-cited scientists based mostly on quotation metrics that incorporate not solely the variety of citations obtained, but in addition info on and adjustment for co-authorship and for authorship positions amongst printed papers.


We used the strategy we’ve got beforehand utilized [1214] for producing a composite quotation index and the development of complete databases together with the two% top-cited scientists in every of 174 scientific subfields, as outlined by the Science-Metrix (RRID:SCR_024471) classification [11]. These subfields cowl all forms of science, know-how, and (bio)medication in addition to scholarly disciplines on the research of humanities and social disciplines. We use summarily the phrases “science,” “scientific fields,” and “scientists” within the paper to cowl all these numerous forms of scholarship, despite the fact that, strictly talking, a few of these authors might not see themselves as “scientists.” Every scientist might have printed papers in additional than 1 subfields, however ultimately he/she is classed right into a single dominant subfield, the one which has the best share amongst his/her papers. The utilized Science-Metrix classification makes use of allocation every journal in a single subfield, apart from multidisciplinary journals the place the articles could also be cut up into a number of subfields. The first purpose of the evaluation was to evaluate throughout every of the 174 scientific subfields to what extent the gender imbalance in illustration of girls among the many top-cited scientists has been ameliorating over time. We will monitor the illustration percentage-wise of girls among the many top-cited scientists with totally different publication ages.

We used NamSor (RRID:SCR_023935) [15], a gender-assignment software program to assign gender to the Scopus (RRID:SCR_022559) creator IDs. We have now beforehand used Scopus information to assign gender to every creator in initiatives executed by the ICSR Lab, such because the European Fee She Figures report: [16] and the Elsevier Gender report [17]. The NamSor software programming interface takes first/final title and nation into consideration. For nation, we took the primary nation an creator publishes from as the perfect estimate (the nation of his/her oldest printed paper). Some authors might transfer to totally different international locations throughout their profession after which there is no such thing as a excellent answer on which nation ought to symbolize them. Nevertheless, gender bias has formative affect ranging from very early age, due to this fact assigning these authors to the nation of their oldest publication might be probably the most acceptable alternative. Info on nation of start (which can even be totally different from the nation of oldest publication) is just not obtainable in Scopus and very tough to seek out for many authors. We solely stored gender assignments with a confidence rating >85%.

We created databases of two% top-cited authors just like those that we’ve got created and up to date on an annual foundation (final up to date based mostly on September 1, 2022 information utilizing info on over 9 million authors with not less than 5 printed full papers throughout all science) [1214]. Briefly, the scientists are ranked based mostly on a composite science indicator that thought-about 6 quotation metrics (whole citations, h-index, co-authorship adjusted hm-index, variety of citations to single-authored papers, variety of citations to single and first-authored papers, and variety of citations to single-, first-, or last-authored papers). The composite indicator thus takes into consideration not solely the general quotation affect, but in addition co-authorship, and particularly the quotation affect from papers the place the creator has had authorship positions that in most fields counsel better contribution to the work.

Within the present venture, as an alternative of contemplating all authors collectively, we thought-about in separate runs:

A. These with first publication earlier than 1992 (30+ years of publication age),

B. 1992 to 2001 (20 to 30 years of publication age),

C. 2002 to 2011 (10 to twenty years of publication age),

D. 2012 or later (10 or fewer years of publication age).

For every of the 4 units, we generated the listing of the two% top-cited scientists in every of the 174 subfields with the rating based mostly on the composite index. We generated information each for career-long affect (all citations obtained at any time for all papers printed at any time) and for the quotation affect in the latest calendar yr, on this case 2021 (citations obtained in 2021 to papers printed in any time). This was executed twice: with and with out together with self-citations (as executed in our earlier work). Outcomes are largely comparable for career-long and single current yr affect; we report intimately right here the latter (single current yr affect) and likewise present the principle outcomes in accordance with the previous (career-wide affect) strategy. Furthermore, the introduced analyses contemplate as top-cited all scientists who’re within the top-2% in accordance with the composite index rating both within the calculations excluding self-citations and/or within the calculations together with self-citations. The overwhelming majority of included scientists are within the top-2% utilizing each calculations. We then estimated the share of ladies and men within the 4 publication age cohorts for every of the scientific subfields.

In calculating whether or not the proportion of girls among the many top-cited scientists modifications over time, we excluded unclear names that can not be assigned to a gender with >85% certainty. Gender task is tougher in names from some international locations than for others and it additionally is dependent upon whether or not the complete first names can be found relatively than simply the primary title preliminary.

We targeted extra on subfields which have reached a share of girls of not less than 50% (matching or outnumbering males) and of not less than 30% and at what age cohorts these milestones had been achieved. Comparatively, we additionally targeted on the different finish of the spectrum, subfields the place the share of girls remained under 10%.

We additionally calculated the relative propensity R of girls versus males to be among the many 2% top-cited in every subfield and publication age cohort. If there are n(w) ladies and n(m) males in a given subfield and given publication age cohort, and N(w) and N(m) amongst them are within the top-2% most-cited, then .

We deal with subfields and publication age cohorts with R> = 1, i.e., the place ladies have a bigger relative illustration among the many top-cited scientists than their illustration within the total depend of authors. Reciprocally, we additionally deal with subfields and publication age cohorts the place R<1/3, i.e., the place males have greater than 3-fold relative overrepresentation amongst top-cited scientists than within the total depend of authors.

All these analyses had been primarily carried out utilizing the worldwide information of all scientists no matter nation. We then rerun these analyses restricted to the scientists who’re from high-income international locations and individually for non-high-income international locations and for scientists who’re from the USA of America. Excessive-income international locations are these categorized as such by the World Financial institution in 2023 (

We additionally current information for the ratio of males over ladies amongst authors and, comparatively, amongst top-cited authors individually for every nation, focusing totally on the youngest (post-2011) cohort to see if some particular international locations have ameliorated imbalances greater than others. Knowledge are introduced for every nation for all scientific subfields mixed (country-level information per subfield have largely very small numbers).

Lastly, we assessed whether or not ladies are deprived in tutorial recruitments and promotions even when their early work has excessive affect. Specializing in the youngest top-cited scientists (post-2011 cohort) who’re nonetheless pretty early of their profession, we chosen a random pattern of 100 males and 100 ladies. We manually checked their info on-line to determine what number of of them are as of April 2023 in academia, business, authorities, or different occupation. All retrieved sources had been eligible for perusal, together with, however not restricted to LinkedIn, ResearchGate, Google Scholar, Frontiers Biosketch pages, and private and CV-related pages in college and different establishment web sites. Amongst those that had been positioned in academia, we recorded what number of of them had been full professors, affiliate professors, or assistant professors (or comparable early-level college title). We coded totally different tutorial titles into these 3 ranks, based mostly on perceived equivalence, e.g., for United Kingdom readers could also be thought-about the equal of affiliate professors. The guide inspection of 100 + 100 = 200 random pattern of scientists was additionally used to look at concurrently whether or not there was a considerable accuracy drawback with Scopus ID assignments, i.e., whether or not top-cited scientists with seemingly quick publication age had been artifacts of older researchers with fragments of their current analysis output separated from their earlier work. This artifact arises as a result of in just a few instances, the publications of an creator are cut up in 2 or extra totally different creator ID information in Scopus that will comprise papers protecting totally different yr spans; e.g., if an creator who has printed from 2000 till now has her publications cut up right into a Scopus ID file that covers the papers printed from 2000 till 2020 and a unique Scopus ID file that covers the papers printed after 2020, that second Scopus ID file will seem as if it belongs to a younger creator. It additionally allowed further validation of the gender task generated by the NamSor algorithm.

It is a descriptive, exploratory evaluation of a science-wide giant bibliometric dataset (not pre-registered). We carried out exploratory statistical testing utilizing evaluation of two × 4 tables adjusting for development with precise take a look at and comparability of proportions. P-values are two-tailed.


Proportion of girls amongst all authors and top-cited authors

Throughout a complete of 9,071,122 authors with not less than 5 full papers, the algorithm assigned 3,784,507 as males, 2,011,616 as ladies, and for 3,274,999 (36.1%), the task to gender was unsure. Among the many top-cited authors, 101,918 had been assigned as males, 31,725 as ladies, and 61,672 had been unsure. Unsure gender authors will not be thought-about in any additional calculations. Total, males outnumbered ladies 1.88-fold amongst all authors and three.21-fold amongst top-cited authors.

As proven in Desk 1, there was growing illustration of girls in cohorts of authors with newer first publication yr each for all authors and for top-cited authors. The ratio of males to ladies for all authors was 3.93 for authors who first printed earlier than 1992 and step by step decreased to 2.06 for authors first publishing in 1992 to 2001, 1.57 for authors first publishing in 2002 to 2011, and 1.36 for authors first publishing after 2011. There was bigger gender inequality for top-cited authors in any respect age teams, with the respective ratios being 6.41, 3.48, 2.74, and a pair of.28. R (the ratio of ratios for top-cited and all authors) remained fixed throughout age cohorts.

In youthful age cohorts, the proportion of girls amongst top-cited authors improved throughout disciplines (Fig 1). Amongst top-cited authors who printed their first paper earlier than 1992, in virtually half of the 174 scientific subfields (n = 69, 40%) ladies represented solely 0% to 10% of the extremely cited authors (i.e., ratio of males to ladies was > = 9). The variety of scientific subfields with such main underrepresentation of girls among the many top-cited authors decreased sharply over time and among the many youngest age cohort (authors with first publication after 2011), solely 3 subfields had such a sample (Desk 1). Within the pre-1992 age cohort, solely 5 subfields had > = 50% illustration of girls among the many top-cited authors and the variety of subfields with > = 50% illustration of girls elevated step by step to achieve 32 (18%) within the post-2011 age cohort. There was a a lot bigger, gradual enhance within the variety of subfields the place ladies represented > = 30% of the top-cited authors, from 22 (13%) within the pre-1992 cohort to 97 (56%) within the post-2011 cohort (Desk 1).

Desk 2 reveals the 32 subfields the place top-cited ladies matched or outnumbered top-cited males within the youngest (post-2011) age cohort. As proven, in all of those 32 subfields with 1 exception (Social Science Strategies), there was a bigger pool of girls than males authors beginning their publications post-2011. The 32 subfields cowl all kinds of disciplines, with heavier concentrations in medication, well being sciences, social sciences, and cultural domains, and distinct absence of mathematical, engineering, and economics subfields. Detailed information on all 174 subfields are positioned in Elsevier Knowledge Repository, doi: 10.17632/wwykk8d48g.3.

The propensity of girls to seek out themselves among the many top-cited (after accounting for variety of whole obtainable authors) exceeded the efficiency of males in 29/174 subfields for the pre-1992 cohort, in 33/174 subfields within the 1992 to 2001 cohort, in 17/174 subfields within the 2002 to 2011 cohort, and in solely 12/174 subfields within the youngest (post-2011) cohort. On the different finish of the spectrum, subfields with greater than 3-fold propensity benefit for males (R<1/3) decreased from 24/179 within the pre-1992 cohort, to 17/174 within the 1992 to 2001 cohort, to 14/179 within the 2002 to 2011 cohort, and 10/174 within the post-2011 cohort. The ten subfields with R<1/3 within the youngest (post-2011) cohort had been Financial Idea, Econometrics, Structure, Microscopy, Music, Common Physics, Paleontology, Biophysics, Speech Language Pathology, and Mechanical Engineering.

Analyses on excessive and non-high-income international locations

Within the total database, there was virtually double the variety of authors from high-income international locations (5,899,402) than from non-high-income international locations (3,171,720) and unsure gender task was much less widespread within the former than within the latter (25.3% versus 56.2%). Males outnumbered ladies extra prominently in high-income international locations (2,925,898/1,481,038 = 1.98) than in different international locations (858,609/530,578 = 1.62) within the variety of authors. The two teams of nations had an identical preponderance of males over ladies among the many top-cited authors (85,776/26,663 = 3.18 and 16,142/4,762 = 3.39, respectively).

Nevertheless, when specializing in the youngest cohort (first publishing after 2011), there was an equal variety of authors from high-income and different international locations (1,259,314 versus 1,248,084) and each teams of nations had an identical ratio of male to feminine authors (576,212/415,307 = 1.36 versus 336,415/250,579 = 1.34), whereas the preponderance of males over ladies among the many top-cited authors had decreased within the high-income international locations, however not considerably within the different international locations (17,742/8,472 = 2.09 versus 7,325/2,499 = 2.93).

As proven in Desk 3, there was a gradual enhance over time within the variety of subfields the place ladies represented > = 50% of the top-cited authors in high-income international locations (from 3% within the pre-1992 cohort to 21% within the post-2011 cohort), however this was not seen in different international locations (nonetheless 6% within the post-2011 cohort). The variety of subfields the place ladies represented = <10% of the top-cited authors decreased for each high-income and different international locations, however 15% of subfields in non-high-income nonetheless confirmed = <10% ladies authors even within the youngest (post-2011) cohort. Detailed information on all 174 subfields seem in Elsevier Knowledge Repository, doi: 10.17632/wwykk8d48g.3.

Nation-level analyses

Totally different international locations different within the extent of imbalance for the ratio of illustration of males versus ladies amongst all authors and amongst top-cited authors. Fig 2 presents these ratios for the youngest (post-2011) cohort for the 52 international locations with greater than 5,000 authors (numbers are small and make these ratios extra unreliable for different international locations). Outcomes had been comparable when top-cited authors had been decided based mostly on career-long affect (Fig 2A) or single current yr affect (Fig 2B) and we focus on right here intimately the latter. Within the youngest cohort, 11 international locations had fewer males than ladies authors (the bottom ratios had been 0.63 in Thailand and 0.87 in Italy), 1 nation had an equal quantity for each genders and 41 had extra males than ladies (the best ratios had been 6.85 in Iraq and 4.06 in Saudi Arabia). Within the youngest cohort, no international locations had fewer males than ladies top-cited authors, however the ratio was closest to 1.00 for Italy (1.03) and Romania (1.04). Conversely, the best ratios of males versus ladies top-cited authors had been seen in Iraq (14.2) and Japan (9.92). India, Colombia, Pakistan, Argentina, Finland, and Japan had the more serious deterioration of the gender imbalance when top-cited authors had been thought-about relatively than all authors (ratio of ratios, 4.95, 4.38, 3.70, 3.36, 2.98, and a pair of.92, respectively). Of those 6 international locations, Argentina and Finland had extra ladies than males authors total, however giant imbalances favoring males amongst top-cited authors. Detailed information on all international locations and age cohorts are in Elsevier Knowledge Repository, doi: 10.17632/wwykk8d48g.3.


Fig 2.

Ratio of males over ladies for all authors (horizontal axis) and ratio of males over ladies for top-cited authors (vertical axis) for the youngest age cohort (authors who began publishing after 2011) for the 52 international locations that had greater than 5,000 authors in that cohort. International locations with fewer authors have too few top-cited authors and the gender ratio for top-cited authors can be pushed by very small numbers. The two panels present outcomes when top-cited authors are decided in accordance with career-long affect and when they’re decided in accordance with single current yr affect. The information underlying this determine may be present in

In-depth guide evaluation of random samples from the youngest cohort

Within the random pattern of 100 authors assigned as ladies by the algorithm and chosen from the post-2011 cohort, on shut guide verification, 3 had been males and 10 had been authors who had truly began publishing earlier than 2011 however their earlier publications had not been assigned by Scopus of their chosen primary creator profile. Among the many respective random pattern of 100 authors assigned as males, 9 had equally began printed in earlier years. Thus, 87 verified eligible ladies and 91 verified eligible males had their work histories evaluated in-depth for his or her present occupations (Desk 4). There have been some attainable developments for extra ladies being engaged in authorities positions than males and extra males than ladies working within the business, however the distinction may have been resulting from likelihood. The bulk in each gender samples had been in a tutorial surroundings; amongst them, barely greater than half already had some tutorial title and this tended to be extra widespread for males than for ladies (44/69 (64%) versus 32/59 (54%)), however the distinction was not statistically vital (p = 0.29). Solely 4 had full professor appointments (2 ladies, 2 males; all 4 in non-high-income international locations).


Our analysis of a complete science bibliometric database with over 9 million authors who’ve printed not less than 5 full papers reveals that there have been substantial corrections of the gender imbalance within the scientific workforce over time. Nevertheless, these corrections are nonetheless lagging behind in lots of scientific subfields and fluctuate extensively throughout international locations. Furthermore, whereas the distinction between the variety of female and male authors has total turn into modest (about 1.3-fold throughout all scientific authors), the distinction within the variety of top-cited authors between the two genders stays a lot greater. The general imbalance on this regard is about 2-fold in high-income international locations (and likewise within the USA particularly) and 3-fold in different international locations. There’s presently very giant heterogeneity throughout scientific subfields and international locations within the presence and prominence of gender imbalances. Within the youngest cohort of scientists (those that began publishing after 2011) in virtually 1 in 5 subfields, ladies match or outnumber males among the many ranks of its top-cited scientists. Nevertheless, in virtually all of those subfields this largely displays the truth that extra ladies than males work and publish in them. Conversely, even within the youngest cohort of scientists, in 44% of the scientific subfields ladies symbolize lower than 30% of the top-cited authors. Lastly, many of the youngest top-cited scientists are nonetheless working in tutorial environments and there was a development for extra males than ladies to have positions within the typical tutorial ladder (assistant, affiliate, or full professor). However, only a few have reached full professor appointments.

We additionally examined relative propensity metrics that appropriate the ratio of top-cited authors by contemplating additionally the ratio of all authors who’re ladies versus males. We famous that over time, there have been fewer subfields the place ladies had a aggressive benefit towards males to seek out themselves among the many top-cited authors as soon as they began publishing in a given subfield. Concurrently, the variety of subfields the place males had a big aggressive benefit to seek out themselves among the many top-cited authors as soon as they began publishing in a given subfield additionally diminished over time.

The scientific workforce globally is altering. There’s a fast creation of huge analysis productiveness in analysis and scientific publications in some non-high-income international locations like China [1820], usually with monetary incentives which have attracted criticism [18]. Due to this fact, the share of authors from non-high-income international locations has elevated sharply. Among the many youngest authors, these with a decade or much less of Scopus-indexed publication historical past, half of them come from non-high-income international locations. In these international locations, gender imbalances within the capability to achieve the top-cited group stay a lot stronger than in high-income international locations. This poses an additional problem to reaching fairness in these international locations, the place analysis is usually carried out beneath suboptimal circumstances. Earlier work has proven that the gender hole in science, know-how, and medication fields is smaller in international locations the place ladies usually tend to main in these fields [21]. Furthermore, there are native elements and obstacles in every low- and middle-income nation that form the traits of its workforce and the extent of distortion from gender bias [22].

Even high-income international locations, together with the USA, have a lot room to optimize fairness. Within the USA, it has been documented that ladies are much less more likely to be included as authors particularly in extremely cited papers [23]. Japan has 10-fold extra top-cited males than ladies even within the youngest cohort, in all probability a mirrored image of long-lasting traditions [24], regardless of efforts to enhance perceived norms [25]. Different international locations comparable to Argentina and Finland have sturdy gender imbalances in top-cited authors, despite the fact that they’ve managed to extinguish imbalances within the total variety of authors.

A lot consideration must be given to the youthful generations of scientists to advertise fairness and optimize their path in analysis. Our in-depth evaluation of a random pattern of scientists who’ve been publishing Scopus-indexed papers for a decade or much less reveals a development for extra males than ladies to have entered the tutorial ladder, though the samples had been fairly small and the distinction might be resulting from likelihood. We must always warning that info obtainable on-line on tutorial ranks will not be all the time up-to-date or full, however any missingness might be not affected by gender. A worrisome commentary is that only a few of those early overachievers have reached professor-level appointments. Not one of the 4 full professors on this pattern got here from a high-income nation and we can’t exclude the likelihood that the 4 full professors had additionally began publishing sooner than 2012 in journals that aren’t listed in Scopus (e.g., native journals) and thus might have longer publication ages. In most tutorial environments and in most international locations, progress by the tutorial ranks takes a painfully very long time and funding independence is often reached within the mid-40s [26]. Funding disparities additionally live on they usually might gas profession selections and development [27]. The present state of affairs could also be eroding impartial creativity and must be challenged [28]. Previously, the time to get a doctoral diploma was shorter and scientists may turn into college and even full professors very quickly after acquiring their doctoral diploma. Prolonged graduate research and a number of postdoctoral experiences are presently much more widespread earlier than reaching independence [2931]. Very gifted people who present clear early promise might have to be promoted a lot quicker. Maybe total analysis originality and creativity will get promoted if younger gifted scientists are given extra assist and confidence in tenure.

We thought-about all authorship positions in counting variety of scientists of various genders. Nevertheless, the calculation of the composite quotation indicator that’s used to determine the top-cited scientists provides numerous weight to single, first, and final authorships, versus center authorship. It’s attainable that in some instances, ladies could also be extra more likely to be listed as secondary or supporting authors relatively than first or senior authors (or not be listed in any respect) [23] and this may affect their visibility and recognition within the tutorial group. Thus, a number of forces might converge in the direction of diminishing the possibilities of ladies turning into top-cited.

Our work has some limitations. First, for over a 3rd of the authors, gender task was unsure and these folks needed to be excluded from additional analyses. This stage of uncertainty is unavoidable with any automated gender task instrument. Unsure gender was modestly much less widespread among the many top-cited scientists and in these from high-income international locations, but it surely was very excessive in authors from non-high-income international locations. Whereas there is no such thing as a motive to imagine that the illustration of males versus ladies can be totally different within the unsure gender group, we can’t exclude the likelihood for some imbalance, e.g., if ladies are extra probably to make use of solely initials relatively than full first title. One ought to due to this fact be cautious concerning the uncertainty that these excluded authors induce in the principle calculations of gender ratios. Furthermore, even with a >85% certainty choice threshold, some folks shall be assigned to the fallacious gender by NamSor. This fallacious task occurred however in solely 6 of 200 randomly chosen authors examined in depth. It’s attainable that the danger of mistaken gender task varies throughout fields and it could be extra more likely to replicate ladies being assigned to male gender than the alternative. In that case, this may occasionally trigger some underestimation of the share of girls in some international locations. Second, Scopus is a complete database, however some forms of publications, e.g., books and a few particular journals will not be represented correctly [10]. This will have an effect on the validity of the rating in some scientific subfields (which particular people are included within the top-cited), however it’s much less more likely to have an effect on gender ratios. Third, we used a beforehand extensively validated methodology for figuring out the top-cited authors. Nevertheless, as we’ve got described earlier than intimately [1214], all quotation metrics and calculations, together with ours, have deficiencies and inaccuracies [32]. Furthermore, quotation affect (in no matter kind it’s calculated) shouldn’t be construed as an ideal surrogate of analysis high quality or real-world affect. However, our strategy provides a reproducible option to determine authors with the best quotation metrics. Imbalances between genders might fluctuate in diploma throughout totally different metrics or points of labor achievement. Lastly, we may solely have a look at the excellence between women and men, a binary classification that doesn’t contemplate self-perceptions of gender which transcend binary choices. It is a recognized unavoidable limitation of any algorithm that tries to assign gender based mostly on names’ and international locations’ info.

Permitting for these caveats, this large-scale evaluation provides insights for previous and present standing of gender imbalances in scientific productiveness and prime quotation affect and could also be used for future planning and analysis. Comparable information may additionally be used for benchmarking inside single international locations and establishments. The persisting giant imbalances in a number of scientific disciplines want extra research to grasp their causes. One may additionally study quite a bit from disciplines the place ladies have matched and even outperformed males in productiveness and quotation affect. The counterfactual of best fairness might not symbolize a state of affairs the place women and men have equal illustration among the many top-cited scientists in each subfield. However, the large composite image suggests that there’s nonetheless substantial room for additional correction of imbalances.


  1. 1.
    Kozlowski D, Larivière V, Sugimoto CR, Monroe-White T. Intersectional inequalities in science. Proc Natl Acad Sci U S A. 2022 Jan 11;119(2):e2113067119. pmid:34983876
  2. 2.
    Ni C, Smith E, Yuan H, Larivière V, Sugimoto CR. The gendered nature of authorship. Sci Adv. 2021 Sep 3;7(36):eabe4639. pmid:34516891
  3. 3.
    Larivière V, Ni C, Gingras Y, Cronin B, Sugimoto CR. Bibliometrics: international gender disparities in science. Nature. 2013 Dec 12;504(7479):211–3. pmid:24350369
  4. 4.
    Eloy JA, Svider PF, Cherla DV, Diaz L, Kovalerchik O, Mauro KM, et al. Gender disparities in analysis productiveness amongst 9952 tutorial physicians. Laryngoscope. 2013 Aug;123(8):1865–75. pmid:23568709
  5. 5.
    Carr PL, Gunn C, Raj A, Kaplan S, Freund KM. Recruitment, Promotion, and Retention of Ladies in Educational Drugs: How Establishments Are Addressing Gender Disparities. Womens Well being Points. 2017 Might-Jun;27(3):374–381. pmid:28063849
  6. 6.
    De Kleijn M, Jayabalasingham B, Falk-Krzesinski HJ, Collins T, Kuiper-Hoyng L, Cingolani I, et al. The Researcher Journey By way of a Gender Lens: An Examination of Analysis Participation, Profession Development and Perceptions Throughout the Globe. (Elsevier, March 2020) Retrieved from
  7. 7.
    Ioannidis JPA, Thombs BD. A person’s information to inflated and manipulated affect elements. Eur J Clin Investig. 2019 Sep;49(9):e13151. pmid:31206647
  8. 8.
    Card D, DellaVigna S, Funk P, Iriberri N. Gender gaps on the academies. Proc Natl Acad Sci U S A. 2023 Jan 24;120(4):e2212421120. pmid:36656862
  9. 9.
    Huang J, Gates AJ, Sinatra R, Barabási AL. Historic comparability of gender inequality in scientific careers throughout international locations and disciplines. Proc Natl Acad Sci U S A. 2020 Mar 3;117(9):4609–4616. pmid:32071248
  10. 10.
    Baas J, Schotten M, Plume M, Côté G, Karimi R. Scopus as a curated, high-quality bibliometric information supply for educational analysis in quantitative science research. Quant Sci Stud. 2020;1:377–386.
  11. 11.
    Archambault E, Beauchesne OH, Caruso J. “In the direction of a multilingual, complete and open scientific journal ontology” in Proceedings of the thirteenth Worldwide Convention of the Worldwide Society for Scientometrics and Informetrics (ISSI), Durban, South Africa. Noyons B, Ngulube P, Leta J, editors. 2011, p. 66–77.
  12. 12.
    Ioannidis JP, Klavans R, Boyack KW. A number of quotation indicators and their composite throughout scientific disciplines. PLoS Biol. 2016;14:e1002501. pmid:27367269
  13. 13.
    Ioannidis JPA, Baas J, Klavans R, Boyack KW. A standardized quotation metrics creator database annotated for scientific area. PLoS Biol. 2019;17:e3000384. pmid:31404057
  14. 14.
    Ioannidis JPA, Boyack KW, Baas J. Up to date science-wide creator databases of standardized quotation indicators. PLoS Biol. 2020;18:e3000918. pmid:33064726
  15. 15.
    NamSor. Accessible from:
  16. 16.
    European Fee She Figures report. Accessible from:, final accessed April 15, 2023.
  17. 17.
    Elsevier Gender report. Accessible from:, final accessed April 15, 2023.
  18. 18.
    Hvistendahl M. China’s publication bazaar. Science. 2013 Nov 29;342(6162):1035–9. pmid:24288313
  19. 19.
    Jones TS, Plume AM. Monitoring China’s publication growth. Nature. 2011 Might 12;473(7346):154. pmid:21562546
  20. 20.
    Li J, Zhu X, Wu D. China’s publications: fewer however higher. Nature. 2021 Apr;592(7855):507. pmid:33879887
  21. 21.
    Aldén and Neuman. Tradition and the gender hole in alternative of main: An evaluation utilizing sibling comparisons. J Econ Behav Organ. 2022.
  22. 22.
    Rose and Hardi. “With Training You Can Face Each Wrestle”: Gendered Greater Training in Iraq and Iraqi Kurdistan—Half Three: The Gender Downside.
  23. 23.
    Ross MB, Glennon BM, Murciano-Goroff R, Berkes EG, Weinberg BA, Lane JI. Ladies are credited much less in science than males. Nature. 2022;608:135–145.
  24. 24.
    Okoshi Ok, Nomura Ok, Fukami Y, Tomizawa Y, Kobayashi Ok, Kinoshita Ok, et al. Gender inequality in profession development for females in Japanese Educational Surgical procedure. Tohoku J Exp Med. 2014;234:221–227. pmid:25355369
  25. 25.
    Nagano N, Watari T, Tamaki Y, Onigata Ok. Japan’s tutorial obstacles to gender equality as seen in a comparability of private and non-private medical faculties: a cross-sectional research. Ladies’s Well being Rep (New Rochelle). 2022 Jan 31;3(1):115–123.
  26. 26.
    Levitt M, Levitt JM. Way forward for elementary discovery in US biomedical analysis. Proc Natl Acad Sci U S A. 2017 Jun 20;114(25):6498–6503. pmid:28584129
  27. 27.
    Lauer MS, Roychowdhury D. Inequalities within the distribution of Nationwide Institutes of Well being analysis venture grant funding. elife. 2021 Sep 3;10:e71712. pmid:34477108
  28. 28.
    Ioannidis JP. Extra time for analysis: fund folks not initiatives. Nature. 2011 Sep 28;477(7366):529–31. pmid:21956312
  29. 29.
    Institute of Drugs. The Postdoctoral Expertise Revisited. Washington, DC: The Nationwide Academies Press; 2014.
  30. 30.
    Andalib MA, Ghaffarzadegan N, Larson RC. The postdoc queue: A labour pressure in ready. Syst Res Behav Sci. 2018;35(6):675–686.
  31. 31.
    Denton M, Borrego M, Knight DB. U.S. postdoctoral careers in life sciences, bodily sciences and engineering: Authorities, business, and academia. PLoS ONE. 2022;17(2):e0263185. pmid:35108316
  32. 32.
    Hicks D, Wouters P, Waltman L, de Rijcke S, Rafols I. Bibliometrics: The Leiden Manifesto for analysis metrics. Nature. 2015 Apr 23;520(7548):429–31. pmid:25903611


Please enter your comment!
Please enter your name here

Most Popular

Recent Comments